Semantic XBRL

Semantic XBRL is a repository that contains semantic data generated from the data submitted by participants in the XBRL Program EDGAR promoted by the U.S Securities and Exchange Commission (SEC). Semantic data is generated by translating XBRL XML to RDF following the XML Semantics Reuse Methodology, which is implemented by the ReDeFer tools. First, there is the XSD2OWL translation, which generates XBRL OWL ontologies from the XBRL XML Schemas. Then XML2RDF, which translate XBRL XML instances to RDF taking into account the XBRL OWL ontologies.

Status

January 2009: 1,03 million triples have been generated from 489 XBRL filings.
April 2009: 1,34 million triples from 612 XBRL filings.
March 2010: 2,98 million triples from 1201 XBRL filings.
June 2010: 4,15 million triples from 1766 filings (dump).

The resulting RDF partially mimics the original XBRL structure from which it is automatically generated. Some small changes are introduced, mainly a new FactType class for facts. Fig. 1 shows a diagram for the core concepts of the resulting Semantic XBRL model. The resulting data is semantically enriched so it is easier to integrate different filings and to cross query filings for different dates, companies, accounting principles,...

Some example queries.

Fig. 1. Semantic XBRL core concepts